Struct CompletionRequest 

pub struct CompletionRequest {
    pub system: Option<String>,
    pub prompt: String,
    pub max_tokens: u32,
    pub temperature: Option<f64>,
    pub reasoning: bool,
    pub effort: Option<Effort>,
    pub thinking_budget: Option<u32>,
    pub json_list: bool,
}

A single-turn completion request. The model is bound to the client.

Fields

system: Option<String>

Optional system prompt.

prompt: String

The user message.

max_tokens: u32

Upper bound on generated tokens, reasoning included.

temperature: Option<f64>

Sampling temperature; the provider default applies when None.

reasoning: bool

Whether to ask the model for reasoning.

effort: Option<Effort>

Reasoning effort, used only when reasoning is true.

thinking_budget: Option<u32>

Fixed extended-thinking token budget, used only when reasoning is true and only by the anthropic protocol client. When set, it takes precedence over effort for that client: the request asks for a fixed thinking budget instead of adaptive thinking, and omits output_config.effort.

json_list: bool

Whether the answer must be a list of strings. The openai protocol client then asks for structured output (a JSON schema); other clients ignore it. The list stages set it so compliant models return parseable JSON.

Implementations

impl CompletionRequest

pub fn for_role( role: &RoleModel, system: Option<String>, prompt: String, ) -> Self

Builds a request with the parameters configured for role.

Trait Implementations

impl Clone for CompletionRequest

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CompletionRequest

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for CompletionRequest

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for CompletionRequest